1. Computer problem? Tech Support Guy is completely free -- paid for by advertisers and donations. Click here to join today! If you're new to Tech Support Guy, we highly recommend that you visit our Guide for New Members.

Opinions on damaged hard drive

Discussion in 'Hardware' started by jiml8, Oct 29, 2011.

Thread Status:
Not open for further replies.
Advertisement
  1. jiml8

    jiml8 Guest Thread Starter

    Joined:
    Jul 2, 2005
    Messages:
    2,634
    My workstation has 5 SCSI drives in it, running off of an Adaptec 29320 controller.

    A couple of weeks ago (2 weeks today, as it happens) I had a SCSI cable short which somehow took out the controller card and burned a ground trace on the onboard controller card on one of the hard drives (/dev/sdd, a 15K RPM 147 GB Hitachi drive), AND caused a head crash on another hard drive (/dev/sdb, a 10K RPM 73 GB Fujitsu which is the oldest drive in the system).

    I replaced the cable, had another controller air-freighted to me, found an fixed the burned ground trace, and was back up. Drive /dev/sdd passed all tests after repairing the ground trace and shows no signs of further problems.

    I did not discover the head crash on /dev/sdb for a few days, until I checked the logs and discovered it had failed self-test. I ran Spinrite on the drive, which took about a day and a half to complete, and Spinrite found extensive damage in one region of the platter. Spinrite recovered some significant portion of the data in the damaged section, but failed to recover all data from a total of 81 blocks, and mapped out a total of 167 bad blocks. Data loss consists of one corrupted song file, which doesn't trouble me a bit.

    The drive now appears to be functioning normally but is now failing self-test due to the high number of bad blocks that have been mapped out; this is being identified as an impending failure.

    Now, if these blocks were just going bad, then that would be right and it would be a clear indication that the drive was failing.

    However, this drive experienced a severe external event which presumably won't be repeated, and apparently the read/write head is OK. Also, apparently, the servos are OK since the drive is working.

    The role this drive plays in my system is non-critical. It holds music, some ebooks, some project backups, and I commonly use it as scratchpad disk. I don't routinely back it up because there is nothing on here that I can't easily replace or that is mission-critical. I have copied the entire contents of this disk to another drive at the present time, just in case.

    Now, this repair has so far cost me about $300...new cables, new controller, and overnight shipping. SCSI hard drives are very expensive.

    So, if you were in my position, would you replace the hard drive? Or would you decide to assume that the nature of what happened means the bad block list isn't going to continue to grow?

    For that matter, is a head crash likely to induce further failures? This sort of thing doesn't happen very often, and I really don't know whether this drive can be trusted going forward or not. For now, it seems to be working OK.

    On the other hand, if I replace it with a larger and faster drive (which will cost about $300 or more), I could define a mission critical role for it because I would kind of like to reorganize some of my virtual machines so that there is more efficient usage of the multiple drives in the system...

    So, if you were me, what would you do?
     
  2. ChRoNo16

    ChRoNo16

    Joined:
    May 29, 2005
    Messages:
    727
    Just sit on the drive for a month or two, if it doesnt get worse, you will have a good spare drive, then you can replace with a big new one anyways, and still do VM's on it.
     
  3. Elvandil

    Elvandil

    Joined:
    Aug 1, 2003
    Messages:
    51,988
    If the bad section is contiguous, try partitioning around it. Then test the partitions that you plan to use once in a while to see if any more bad blocks appear. I had a drive that worked for years after doing that.

    But Spinrite can't repair anything, so any blocks that it "recovered" will be back soon. Bad blocks are physical damage to the platter.
     
  4. jiml8

    jiml8 Guest Thread Starter

    Joined:
    Jul 2, 2005
    Messages:
    2,634
    So the consensus is "keep using it".

    I think I'll do that, and just keep monitoring it for health. There's no need to partition around bad blocks; spinrite mapped a bunch of 'em out, and I ran some filesystem tools on the system to tell the filesystem to avoid certain blocks as well. I actually think that the replacement block list for the drive is used up, and therefore not all blocks CAN be mapped out at the firmware level, but I'm not sure of this. Nonetheless, I can map out as many as I need to map out at the filesystem level.

    The damaged area is physically contiguous, apparently representing a crash while the head was moving in the radial direction, meaning that several cylinders are affected in about the same radial location (of course, the platter was spinning, so the actual damage is a section of a spiral, and this actually shows in the pattern of bad blocks).

    I'm contemplating a gradual move to SAS and/or SATA/SSD anyway, so maybe if I must replace this drive I'll replace it with a different technology.
     
  5. jiml8

    jiml8 Guest Thread Starter

    Joined:
    Jul 2, 2005
    Messages:
    2,634
    Well, the situation resolved itself. The drive is showing more bad blocks. Head damage, I suppose.

    Oh well. I've been studying on it for the last week and decided to replace it with another SCSI drive. Other than SSD, they still have the highest performance.

    So I have a 300 GB 15K RPM Fujitsu on the way.
     
  6. Sponsor

As Seen On
As Seen On...

Welcome to Tech Support Guy!

Are you looking for the solution to your computer problem? Join our site today to ask your question. This site is completely free -- paid for by advertisers and donations.

If you're not already familiar with forums, watch our Welcome Guide to get started.

Join over 733,556 other people just like you!

Loading...
Thread Status:
Not open for further replies.

Short URL to this thread: https://techguy.org/1024559

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice